CONCLUSION


In this chapter, we have proposed the concept of social coordination in daily life, which is a mutual concession mechanism for social resources, e.g., space, time, and reservations , through automatic negotiation among software agents rather than through the explicit and verbal communication of human users.

We have also proposed a new kind of architecture, called CONSORTS, for ubiquitous agents in which mass user support services are provided in addition to conventional personal supports.

As an example of social coordination, we have proposed, formalized and analyzed mass user navigation. Although mass user navigation seems to be a planning or scheduling problem at first sight, we have pointed out that conventional problem-solving mechanisms, such as planning, scheduling, GA, reinforcement learning, or stochastic distribution itself, do not work well. To solve the problem, we have proposed a method based on the generation, connection and evaluation of plans, with plan modifications by stochastic distribution and market/auction mechanism, i.e., a kind of bulletin board where users' intentions and preferences are exchanged among users.

Social coordination is not a part of social collaboration. It requires realtime response, although it cannot necessarily generate the best solutions. Realtime response does not seem to be crucial in the theme park problem, but it is really important in other applications, such as social coordination in traffic control, because we do not have much time for decision making in traffic and for navigation guidance when we drive a car. In addition, if we can reduce the amount of traffic in a city or country by just one percent, it will bring much benefit to the economy and environment.

Social coordination is working as an underlying mechanism in our daily lives. Our intention is to enhance such mutual concession mechanisms in a sophisticated way by using software agent technologies. Because this research is just beginning, we will examine and refine the definition of the problem and the algorithm to solve it, first by multi-agent simulation and later by applying it to real situations.




(ed.) Intelligent Agents for Data Mining and Information Retrieval
(ed.) Intelligent Agents for Data Mining and Information Retrieval
ISBN: N/A
EAN: N/A
Year: 2004
Pages: 171

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net